The notebook uses three methods to extract the center and other relevant info about the drop in the image. These methods are:
We have also made some preliminary analysis for all the three methods.
All the code can be found on github repository, specifically on the folder: codes
Codes still need refactoring!
Run on Google Colab
View on Githubimport matplotlib.pyplot as plt
from misc_tools import *
from run import Run
plt.rcdefaults()
# Image Directory
IMAGE_DIRECTORY = "../data/images1"
DATA_DIRECTORY = "../data/results"
r = Run(IMAGE_DIRECTORY, DATA_DIRECTORY)
# Strating Crop coordinates
crop_1 = (150, 40, 40, 40)
crop_2 = (120, 250, 50, 50)
crop_3 = (250, 230, 40, 40)
# Image Positions
positions_1 = (0, 70)
positions_2 = (75, 100)
positions_3 = (106, 155)
df_dc = r.dynamic_cropping(
crop_1,
crop_2,
crop_3,
positions_1,
positions_2,
positions_3,
save=True, #Saving the generated DataFrame
file_name="dc",
plot=False, #whether to plot
strict=True, #Using the equation of line
crop_included=False, #Excluding the crop coordinates
)
Starting dynamic cropping... First step... Second step... Third step... Saving to csv... Done!
get_samples(len_samples = 25,df_path=df_dc,
file_name="using_dc", title="Using DC")
#Smoothing the data
df_dc = smoothen(df_dc)
df_dc
| id | x | y | r1 | r2 | r | vy | vx | v | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | ../data/images1/0.jpg | 870.587824 | 401.000000 | 10.0 | 13.0 | 11.0 | NaN | NaN | NaN |
| 1 | ../data/images1/1.jpg | 870.618264 | 401.301569 | 10.0 | 13.0 | 11.0 | NaN | NaN | NaN |
| 2 | ../data/images1/2.jpg | 870.639818 | 401.579694 | 10.0 | 13.0 | 11.0 | NaN | NaN | NaN |
| 3 | ../data/images1/3.jpg | 870.652864 | 401.810550 | 10.0 | 13.0 | 11.0 | NaN | NaN | NaN |
| 4 | ../data/images1/4.jpg | 870.657780 | 402.039705 | 10.0 | 13.0 | 11.0 | NaN | NaN | NaN |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 139 | ../data/images1/150.jpg | 1160.746077 | 576.424655 | 11.0 | 8.0 | 9.0 | 1.916808 | 4.826432 | 5.347604 |
| 140 | ../data/images1/151.jpg | 1165.276423 | 577.477595 | 11.0 | 8.0 | 9.0 | 2.202874 | 4.853091 | 5.446056 |
| 141 | ../data/images1/152.jpg | 1169.841369 | 578.584610 | 11.0 | 8.0 | 9.0 | 2.435243 | 4.888637 | 5.569656 |
| 142 | ../data/images1/153.jpg | 1174.439592 | 579.800151 | 11.0 | 8.0 | 9.0 | 2.751182 | 4.963698 | 5.715468 |
| 143 | ../data/images1/154.jpg | 1179.000000 | 581.000000 | 11.0 | 8.0 | 9.0 | 3.000000 | 5.000000 | 5.830952 |
119 rows × 9 columns
cols = ["x", "y", "v"]
plot_all(df_dc, title="x, y and v using Dynamical Cropping", cols = cols)
There are some missing values between the timestamps 60 and 80.
cols = ["vx", "vy", "v"]
plot_all(df_dc, title="vx, vy and v using Dynamical Cropping", cols = cols)
cols = ["r1", "r2", "r"]
plot_all(df_dc, title="r1, r2 and r using Dynamical Cropping", cols = cols)
df_si = r.subtracting_images(subtract=True, strict=False,
plot=False)
Getting list of images... Extracting data from images... Error on: ../data/images1/71.jpg Error on: ../data/images1/102.jpg Error on: ../data/images1/160.jpg Error on: ../data/images1/161.jpg Error on: ../data/images1/162.jpg Error on: ../data/images1/163.jpg Error on: ../data/images1/164.jpg Error on: ../data/images1/165.jpg Error on: ../data/images1/166.jpg Error on: ../data/images1/167.jpg Error on: ../data/images1/168.jpg Error on: ../data/images1/169.jpg Error on: ../data/images1/170.jpg Saving to csv... Done!
There are about a dozen of images where the method is not working.
get_samples(len_samples = 25,df_path=df_si,
file_name="using_si", title="Using SI")
#Smoothing the data
df_si = smoothen(df_si)
df_si
| id | x | y | r1 | r2 | r | vy | vx | v | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | ../data/images1/0.jpg | 870.497069 | 401.000000 | 10.000000 | 11.731707 | 10.646436 | NaN | NaN | NaN |
| 1 | ../data/images1/1.jpg | 870.496502 | 401.347136 | 10.052940 | 11.732085 | 10.643978 | NaN | NaN | NaN |
| 2 | ../data/images1/2.jpg | 870.525808 | 401.677822 | 10.105880 | 11.764228 | 10.636604 | NaN | NaN | NaN |
| 3 | ../data/images1/3.jpg | 870.546228 | 402.006239 | 10.127812 | 11.820382 | 10.624315 | NaN | NaN | NaN |
| 4 | ../data/images1/4.jpg | 870.558140 | 402.307052 | 10.157497 | 11.823029 | 10.607109 | NaN | NaN | NaN |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 149 | ../data/images1/151.jpg | 1170.823218 | 579.191151 | 9.607865 | 10.217432 | 9.536774 | 1.020798 | 4.923426 | 5.074575 |
| 150 | ../data/images1/152.jpg | 1175.376631 | 580.361505 | 9.594252 | 10.207790 | 9.493099 | 1.085649 | 4.994706 | 5.146791 |
| 151 | ../data/images1/153.jpg | 1179.918510 | 581.539232 | 9.586689 | 10.219134 | 9.449991 | 1.112876 | 5.033844 | 5.190442 |
| 152 | ../data/images1/154.jpg | 1184.447154 | 582.745131 | 9.616185 | 10.189072 | 9.407071 | 1.140480 | 5.071469 | 5.232244 |
| 153 | ../data/images1/155.jpg | 1189.000000 | 584.000000 | 9.612970 | 10.226130 | 9.403101 | 1.206466 | 5.177349 | 5.338052 |
128 rows × 9 columns
cols = ["x", "y", "v"]
plot_all(df_si, title="x, y and v using Subtracting Images", cols = cols)
cols = ["vx", "vy", "v"]
plot_all(df_si, title="vx, vy and v using Subtracting Images", cols = cols)
cols = ["r1", "r2", "r"]
plot_all(df_si, title="r1, r2 and r using Subtracting Images", cols = cols)
df_ap = r.all_points(plot=False, crop_included=True)
Getting list of images... Extracting data from images... Error on: ../data/images1/71.jpg Error on: ../data/images1/72.jpg Error on: ../data/images1/102.jpg Error on: ../data/images1/159.jpg Error on: ../data/images1/160.jpg Error on: ../data/images1/161.jpg Error on: ../data/images1/162.jpg Error on: ../data/images1/163.jpg Error on: ../data/images1/164.jpg Error on: ../data/images1/165.jpg Error on: ../data/images1/166.jpg Error on: ../data/images1/167.jpg Error on: ../data/images1/168.jpg Error on: ../data/images1/169.jpg Error on: ../data/images1/170.jpg Saving to csv... Done!
Again, there are about a dozen of images where the method is not working.
get_samples(len_samples = 25,df_path=df_ap,
file_name="using_ap", title="Fitting Ellipse")
#Smoothing the data
df_ap = smoothen(df_ap)
df_ap
| id | x | y | r1 | r2 | theta | r | vy | vx | v | |
|---|---|---|---|---|---|---|---|---|---|---|
| 0 | ../data/images1/0.jpg | 869.000000 | 403.000000 | 8.000000 | 11.239176 | 0.251548 | 9.969370 | NaN | NaN | NaN |
| 1 | ../data/images1/1.jpg | 869.070524 | 403.147287 | 8.213651 | 11.240499 | 0.330962 | 9.969749 | NaN | NaN | NaN |
| 2 | ../data/images1/2.jpg | 869.179429 | 403.278881 | 8.425789 | 11.244470 | 0.340577 | 9.970883 | NaN | NaN | NaN |
| 3 | ../data/images1/3.jpg | 869.256192 | 403.440726 | 8.603895 | 11.313103 | 0.403795 | 9.972774 | NaN | NaN | NaN |
| 4 | ../data/images1/4.jpg | 869.331064 | 403.639629 | 8.847230 | 11.306863 | 0.403282 | 9.975421 | NaN | NaN | NaN |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 150 | ../data/images1/153.jpg | 1180.119304 | 582.475137 | 9.773870 | 10.373417 | 1.199565 | 9.943846 | 1.487238 | 4.511817 | 4.764998 |
| 151 | ../data/images1/154.jpg | 1184.573076 | 583.823785 | 9.758366 | 10.364719 | 1.223732 | 9.903573 | 1.613916 | 4.504065 | 4.755489 |
| 152 | ../data/images1/155.jpg | 1189.052562 | 585.211760 | 9.714124 | 10.419739 | 1.234078 | 9.896956 | 1.776139 | 4.494990 | 4.774608 |
| 153 | ../data/images1/156.jpg | 1193.525808 | 586.595198 | 9.741917 | 10.437323 | 1.237128 | 9.892985 | 1.902628 | 4.492721 | 4.800438 |
| 154 | ../data/images1/157.jpg | 1198.000000 | 588.000000 | 9.701834 | 10.495368 | 1.213515 | 9.891662 | 2.000000 | 4.528266 | 4.834646 |
137 rows × 10 columns
cols = ["x", "y", "v"]
plot_all(df_ap, title="x, y and v by Fitting Ellipse", cols = cols)
cols = ["vx", "vy", "v"]
plot_all(df_si, title="vx, vy and v by Fitting Ellipse", cols = cols)
cols = ["r1", "r2", "r"]
plot_all(df_si, title="r1, r2 and r by Fitting Ellipse", cols = cols)
